Don't create redundant unique indexes on PG and SQLite

[rumpuslabs]

PR Info

• Closes redundant unique index on entity-first approach #2873

Bug Fixes

• Prevent creation of redundant UNIQUE indexes on Postgres and SQLite

[tyt2y3]

thank you for the fix! it seems correct, can you briefly explain the root cause of the issue and the fix?

why is this applicable to SQLite as well?

[rumpuslabs]

Postgres and SQLite both automatically create indexes for UNIQUE constraints, so the additional explicit index creation is redundant.

Postgres:

Adding a unique constraint will automatically create a unique B-tree index on the column or group of columns listed in the constraint.

SQLite:

In most cases, UNIQUE and PRIMARY KEY constraints are implemented by creating a unique index in the database.

[Huliiiiii]

It’d help to add a comment here to explain why we filter pg/sqlite out.

[rumpuslabs]

In the process of writing the requested comment, I got to thinking and did some checking, and it seems like MySQL also automatically creates an index from a UNIQUE constraint. I'm not sure where I got the idea that this was a MySQL-specific workaround.

Perhaps the correct fix is simply to remove || column_def.unique from that line? It was added in bc0f6b0 by @tyt2y3

[rumpuslabs]

Or even if column_def.indexed && !column_def.unique since the indexed attribute is redundant?

[tyt2y3]

Or even if column_def.indexed && !column_def.unique since the indexed attribute is redundant?

let's do this. I will make the change